We don't provide the AI.
We provide the questions worth asking it.

You are advising the leadership of a UK contracting business. Attached is a CSV of anonymous, aggregated signal data from our early-warning instrument: daily positive/negative counts per criterion, per function, per project. It contains no individuals and none can be inferred — analyse patterns only. The situation: [FUNCTION] reads amber or worse across ALL of our projects simultaneously. A uniform pattern like this usually indicates a structural cause rather than local ones. Task: 1. Verify the pattern from the data: is the decline genuinely uniform across projects, or is one instance dragging perception? 2. List the plausible structural causes for a uniform decline in this function in this industry, ordered by likelihood, and say which criteria in the data support or weaken each. 3. Distinguish causes that indicate a MISSING OWNER (no one person accountable for this function company-wide) from causes that indicate a shared external dependency (one supplier, one system, one process). 4. Give me five specific, open questions to ask in my next leadership meeting — questions that surface the cause without assigning blame. 5. State plainly what the data CANNOT tell us and what evidence to gather on foot. Be direct. Do not soften findings. If the honest answer is "this function has no owner," say so.

You are a seasoned operations adviser. Attached is a CSV of anonymous aggregate signal data: daily positive/negative counts per criterion within [FUNCTION] on [PROJECT], plus the same function on our other projects for comparison. No individuals are in the data or inferable from it. The situation: this function has been declining for roughly [N] weeks. The lamp has not yet gone red. I need to decide whether this is noise, a temporary dip, or the early stage of a real failure. Task: 1. Characterise the decline: steady or accelerating; broad across criteria or concentrated in one or two? Name the criteria doing the damage. 2. Compare against the same function on the other projects: is this local or the leading edge of something wider? 3. Give your honest read: NOISE / DIP / DETERIORATING, with your confidence and the single piece of data that would most change your mind. 4. If deteriorating: what does this failure mode typically look like on the ground two to four weeks from now if nothing is done? 5. Give me a plan for one working day — who to talk to (by role, not name), what to look at, and the two questions most likely to reveal the cause. I will walk the floor either way. Your job is to sharpen where I look, not to replace the walk.

You are writing for a company board. Attached is a CSV of anonymous aggregate signals from our organisational early-warning instrument, covering the last [N] weeks across all functions and projects. The data contains no individuals. Produce a one-page narrative, British English, plain and unspun: 1. THE PICTURE — two sentences on the overall shape: where signal is strong, where it is thin (functions still below reporting threshold), and what changed since last period. 2. ATTENTION — the two or three function/project instances that most deserve board attention, each with: the trend in one sentence, what the criteria suggest is driving it, and what management is being asked (or should be asked) about it. 3. QUIET HEALTH — one sentence naming what is solidly green, so good work is seen. 4. WHAT THE DATA CANNOT SAY — two sentences of honest limits: these are perception signals from the people closest to the work, not audited facts; they say where to look, not what is true. Rules: no scores or averages across functions — describe distributions and name instances. No praise inflation, no alarm inflation. If something is ambiguous, say it is ambiguous. Maximum 400 words.